In [38]:
In [39]:
In [40]:
For training data , found 0 in folder .DS_Store
For training data , found 15 in folder face1
For training data , found 20 in folder face10
For training data , found 16 in folder face11
For training data , found 14 in folder face12
For training data , found 13 in folder face13
For training data , found 12 in folder face14
For training data , found 15 in folder face15
For training data , found 17 in folder face16
For training data , found 15 in folder face2
For training data , found 14 in folder face3
For training data , found 17 in folder face4
For training data , found 16 in folder face5
For training data , found 16 in folder face6
For training data , found 14 in folder face7
For training data , found 14 in folder face8
For training data , found 16 in folder face9
In [41]:
For testing data , found 0 in folder .DS_Store
For testing data , found 4 in folder face1
For testing data , found 4 in folder face10
For testing data , found 4 in folder face11
For testing data , found 4 in folder face12
For testing data , found 4 in folder face13
For testing data , found 4 in folder face14
For testing data , found 4 in folder face15
For testing data , found 4 in folder face16
For testing data , found 4 in folder face2
For testing data , found 4 in folder face3
For testing data , found 4 in folder face4
For testing data , found 4 in folder face5
For testing data , found 4 in folder face6
For testing data , found 4 in folder face7
For testing data , found 4 in folder face8
For testing data , found 4 in folder face9
In [42]:
Out[42]:
(311, 311, 3)    9
(316, 316, 3)    8
(323, 323, 3)    8
(313, 313, 3)    8
(326, 326, 3)    8
                ..
(301, 301, 3)    1
(338, 338, 3)    1
(278, 278, 3)    1
(342, 342, 3)    1
(368, 368, 3)    1
Length: 83, dtype: int64
In [43]:
Out[43]:
(299, 299, 3)    4
(313, 313, 3)    4
(310, 310, 3)    3
(307, 307, 3)    3
(325, 325, 3)    2
(298, 298, 3)    2
(305, 305, 3)    2
(323, 323, 3)    2
(312, 312, 3)    2
(322, 322, 3)    2
(341, 341, 3)    2
(303, 303, 3)    2
(336, 336, 3)    2
(311, 311, 3)    2
(316, 316, 3)    2
(288, 288, 3)    2
(315, 315, 3)    2
(354, 354, 3)    1
(271, 271, 3)    1
(353, 353, 3)    1
(317, 317, 3)    1
(287, 287, 3)    1
(309, 309, 3)    1
(335, 335, 3)    1
(327, 327, 3)    1
(350, 350, 3)    1
(329, 329, 3)    1
(300, 300, 3)    1
(320, 320, 3)    1
(368, 368, 3)    1
(346, 346, 3)    1
(308, 308, 3)    1
(318, 318, 3)    1
(358, 358, 3)    1
(262, 262, 3)    1
(345, 345, 3)    1
(326, 326, 3)    1
(319, 319, 3)    1
(277, 277, 3)    1
(343, 343, 3)    1
(333, 333, 3)    1
dtype: int64
In [44]:
In [58]:
In [59]:
we have 244 items in X_train
In [60]:
In [61]:
In [62]:
we have 64 items in X_test
In [63]:
In [64]:
X_train shape  is (244, 100, 100, 3)
X_test shape  is (64, 100, 100, 3)
y_train shape  is (244,)
y_test shape  is (64,)
In [65]:
In [66]:
In [67]:
Model Details are : 
Model: "sequential_3"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 conv2d_9 (Conv2D)           (None, 96, 96, 32)        2432      
                                                                 
 max_pooling2d_6 (MaxPooling  (None, 48, 48, 32)       0         
 2D)                                                             
                                                                 
 conv2d_10 (Conv2D)          (None, 44, 44, 100)       80100     
                                                                 
 max_pooling2d_7 (MaxPooling  (None, 22, 22, 100)      0         
 2D)                                                             
                                                                 
 flatten_3 (Flatten)         (None, 48400)             0         
                                                                 
 dense_8 (Dense)             (None, 100)               4840100   
                                                                 
 dense_9 (Dense)             (None, 16)                1616      
                                                                 
=================================================================
Total params: 4,924,248
Trainable params: 4,924,248
Non-trainable params: 0
_________________________________________________________________
None
In [68]:
Epoch 1/10
8/8 [==============================] - 2s 265ms/step - loss: 207.9410 - accuracy: 0.1148
Epoch 2/10
8/8 [==============================] - 2s 261ms/step - loss: 1.4982 - accuracy: 0.5615
Epoch 3/10
8/8 [==============================] - 2s 273ms/step - loss: 0.8125 - accuracy: 0.8033
Epoch 4/10
8/8 [==============================] - 2s 265ms/step - loss: 0.2606 - accuracy: 0.9344
Epoch 5/10
8/8 [==============================] - 2s 265ms/step - loss: 1.0834 - accuracy: 0.7828
Epoch 6/10
8/8 [==============================] - 2s 268ms/step - loss: 0.3862 - accuracy: 0.8852
Epoch 7/10
8/8 [==============================] - 2s 273ms/step - loss: 0.2499 - accuracy: 0.9262
Epoch 8/10
8/8 [==============================] - 2s 267ms/step - loss: 0.5664 - accuracy: 0.8525
Epoch 9/10
8/8 [==============================] - 2s 281ms/step - loss: 0.2212 - accuracy: 0.9508
Epoch 10/10
8/8 [==============================] - 2s 260ms/step - loss: 0.0467 - accuracy: 0.9836
In [69]:
2/2 [==============================] - 1s 61ms/step - loss: 0.1510 - accuracy: 0.9219
Test Loss is 0.1509629786014557
Test Accuracy is 0.921875
In [70]:
In [71]:
In [ ]: